Timezone isn't accessible, please provide the correct parameters
eventFeedUrl=http://realintelligence.com/customers/expos/00Do0000000aAt2/FMS_xmlcreator/a0J5c00001MW1eJ_specific-event-list.xml
trackCategory=Session
eventID=a0J5c00001MW1eJ
timezone=
duration=PTH
, NaNth
1:25-2:30 PM
SARC-302-1: Computational Storage Solutions (System Architectures Track)
Paper Title: Applying Adaptive Computing to Accelerate Computational Storage

Paper Abstract: Computational Storage, the concept of moving compute closer to data and leveraging parallelism to reduce system bottlenecks, is gaining momentum. Further, the interface to computational storage engines is being standardized, thanks to NVMe TP-4091. AMD’s Adaptive Compute Acceleration Platform (ACAP) is a natural fit for deploying this type of technology, owing to ACAP’s flexibility and performance. ACAP’s flexibility is evident when creating a design for compression/decompression. A user can tailor a solution to their workload, choosing various compression algorithms, block size, encoding schemes, and mix of the amount of compression/decompression bandwidth supported. The same flexibility maintains a path for upgrades to evolving standards. In this session, we will provide an example of how a computational storage application for data compression/decompression is detailed, created with the AMD Vitis Unified Software Platform and High-Level Synthesis (HLS). Performance benchmarking results will also be provided.

Paper Author: Seong Kim, Sr. Director Datacenter Architect, AMD

Author Bio: